1 |
On 22 Apr 2016 03:44, Leno Hou wrote: |
2 |
> 2) The problem is due to the wrong $LD="ld -m elf64ppc" when link. |
3 |
> This patch sets $LD="ld -m elf64lppc" in m4/libtool.m4 on ppc64le. |
4 |
|
5 |
i'm looking at the upstream libtool project and i'm not seeing |
6 |
these changes in there. have you sent patches to them to update |
7 |
libtool ? otherwise, all packages will continue to be created |
8 |
with bad versions in them ... |
9 |
|
10 |
what upstream has right now is: |
11 |
if AC_TRY_EVAL(ac_compile); then |
12 |
case `/usr/bin/file conftest.o` in |
13 |
*32-bit*) |
14 |
case $host in |
15 |
... |
16 |
powerpc64le-*linux*) |
17 |
LD="${LD-ld} -m elf32lppclinux" |
18 |
;; |
19 |
powerpc64-*linux*) |
20 |
LD="${LD-ld} -m elf32ppclinux" |
21 |
;; |
22 |
... |
23 |
*64-bit*) |
24 |
case $host in |
25 |
... |
26 |
powerpcle-*linux*) |
27 |
LD="${LD-ld} -m elf64lppc" |
28 |
;; |
29 |
powerpc-*linux*) |
30 |
LD="${LD-ld} -m elf64ppc" |
31 |
;; |
32 |
... |
33 |
|
34 |
your 2.4.2 patch makes sense in that it had no LE support, so you |
35 |
need to add the tuple matching. however, the 2.4.4 & 2.4.6 patches |
36 |
are applying to libtool versions that know how to handle LE. |
37 |
|
38 |
keep in mind that the logic here is just for handling mismatch in |
39 |
the expected bitness and the tuple. i.e. when you're using a tuple |
40 |
like powerpc-linux-gnu (which is normally 32bit) but you're actually |
41 |
producing 64-bit code. |
42 |
-mike |